Behind the Signal: Doral, FL
What’s up with Doral?
Let’s say you’re going through orders, and you come across one with a high order value where the billing and shipping addresses don’t match. You decide to do a bit of sleuthing, starting with research on the shipping city: Doral, FL.
At first glance, shipping to Doral seems like a no-brainer:
- Incomes are high; the average household income is $87K
- It’s the home of a 90-hole Golf Resort and Spa owned by the (Donald) Trump
Based on that information, it’d be perfectly reasonable to ship that order.
However, there’s also cause for caution. Sift Science has found that — despite Doral’s wealth and status as member of the Trump empire — orders shipped there are 8X more risky than normal!
What Versus Why
At Sift, insights like these are discovered automatically, and often the signals are subtle and not immediately intuitive. After all, a computer can say “what”, but it takes a human being to say “why”.
For Doral specifically, I did ask “why”, and here’s what I found.
Doral’s land zoning looks like this:
Yes, Doral is home to not only 90 holes of golf, but also a lot of industrial land!
It has over 3K logistics-related companies and the Miami Free Zone, which offers 750K sq. ft. of duty-free warehouse space. Its proximity to Miami International Airport (the #1 airport in international freight) and Port Miami (which moved 12.5B tons of cargo last year) means it has a thriving logistics industry.
Some of these warehouses offer package forwarding as a service. So, shipping to Doral is risky because that package has a higher likelihood of being forwarded to someplace else.
So what do you do?
Clearly, you shouldn’t blacklist Doral, since most of its inhabitants (population: 48K and growing) could be great customers. Similarly, you shouldn’t blacklist forwarding addresses since not all fraudsters use forwarding addresses and not all forwarding addresses ship to fraudsters.
Ultimately, shipping city is only one factor you should take into account when assessing fraud risk. It could be worth cancelling the order if there were other risky signals, such as how the email was typed in, what products — colors, sizes, etc. — make up the order, and shipping address.
At Sift, we have a proprietary database of over tens of thousands of known risky addresses. However, the number of risky addresses is growing every day. As some addresses get blacklisted by larger retailers, package forwarding companies will change their warehouse locations. Plus, the freight industry itself has been under pressure, and package forwarding as a side-business could be a nice source of incremental profit.
While we’re not experts in package forwarding and freight, we are experts in identifying risk signals associated with this industry using machine learning. We track addresses and more to predict the likelihood of fraud. Follow us on Twitter to learn about more fraud signals or let us know about fraud signals you’ve investigated in the comments below. We’d love to hear them!
